pasys pareto¶

Extract and display the Pareto frontier.

Synopsis¶

pasys pareto <results> -x <metric> -y <metric> [options]

Description¶

The pareto command extracts Pareto-optimal designs from results, displays the top designs, and optionally generates a plot.

Arguments¶

Argument	Required	Description
`results`	Yes	Path to results file (Parquet or CSV)

Options¶

Option	Short	Required	Default	Description
`-x`		Yes		X-axis metric (minimized)
`-y`		Yes		Y-axis metric (maximized)
`--output`	`-o`	No	None	Save Pareto front to file
`--plot`		No	False	Generate plot

Objective Directions¶

By default: - X-axis: Minimized (e.g., cost) - Y-axis: Maximized (e.g., performance)

This covers the common "minimize cost, maximize performance" trade-off.

Examples¶

Basic Extraction¶

pasys pareto results.parquet -x cost_usd -y eirp_dbw

Output:

Pareto Frontier: 12 designs
======================================================================
  case_00023: cost_usd=18234.50, eirp_dbw=38.21
  case_00045: cost_usd=23567.00, eirp_dbw=42.15
  case_00067: cost_usd=31245.00, eirp_dbw=45.67
  case_00012: cost_usd=42890.00, eirp_dbw=48.23
  case_00089: cost_usd=56123.00, eirp_dbw=51.45
  ...

Save to File¶

pasys pareto results.parquet -x cost_usd -y eirp_dbw -o pareto.csv

Creates pareto.csv with all Pareto-optimal designs.

Generate Plot¶

pasys pareto results.parquet -x cost_usd -y eirp_dbw --plot

Creates pareto_plot.png showing: - All designs as blue points - Pareto-optimal designs as red points - Pareto frontier line

Save Plot with Custom Name¶

pasys pareto results.parquet -x cost_usd -y eirp_dbw -o pareto.csv --plot
# Creates: pareto.csv and pareto.png

Common Metric Combinations¶

Cost vs. Performance¶

# Cost vs. EIRP (communications)
pasys pareto results.parquet -x cost_usd -y eirp_dbw

# Cost vs. SNR margin
pasys pareto results.parquet -x cost_usd -y link_margin_db

# Cost vs. detection range (radar)
pasys pareto results.parquet -x cost_usd -y snr_margin_db

Power vs. Performance¶

# Power vs. EIRP
pasys pareto results.parquet -x prime_power_w -y eirp_dbw

Size vs. Performance¶

# Element count vs. EIRP
pasys pareto results.parquet -x n_elements -y eirp_dbw

Output Format¶

Display¶

Shows top 10 Pareto designs with: - Case ID - X metric value - Y metric value

CSV Export¶

Full DataFrame with all columns for Pareto-optimal designs: - All input parameters - All computed metrics - Verification results - Pareto rank and score

Plot¶

PNG image with: - All designs (blue circles) - Pareto-optimal designs (red circles) - Connecting line through Pareto front - Axis labels - Grid

Workflow Example¶

# 1. Run DOE
pasys doe config.yaml -n 200 -o ./study

# 2. Extract Pareto frontier
pasys pareto study/results.parquet -x cost_usd -y eirp_dbw --plot -o study/pareto.csv

# 3. View results
cat study/pareto.csv | head -20
open study/pareto.png  # macOS

# 4. Alternative trade-offs
pasys pareto study/results.parquet -x prime_power_w -y eirp_dbw --plot

Filtering¶

The Pareto extraction uses the verification.passes column to filter feasible designs before extracting the frontier. Only designs that pass all "must" requirements are considered.

Error Handling¶

# File not found
pasys pareto missing.parquet -x cost_usd -y eirp_dbw
# Error: Results file not found: missing.parquet

# Column not found
pasys pareto results.parquet -x invalid_metric -y eirp_dbw
# Error: Column 'invalid_metric' not found in results